Tài liệu GSM, cdmaOne and 3G systems P6 pdf

The maximum user data rate available on a singlephysical channel has since been increased to 14.4 kb/s by reducing the power of the channelcoding on the full rate traffic channel by means

No 3G systems are currently deployed, although trials are in progress As a consequence,this chapter, which deals with systems that are about to be deployed, is treated in a qual-itative manner, describing how they will work rather than quantifying their performances.Before getting into detail, let us briefly review how cellular communications arrived at to-day’s position.

6.1.1 The generation game

There is no doubt that there was pent-up demand for public mobile telephony networks,and when they arrived in the 1980s as the so-called first generation (1G) analogue cellularnetworks, they grew at phenomenal rates These networks initially offered only telephony,but the un-tethering of people from their fixed phones meant that they and businesses couldoperate in completely new ways The Europeans identified in the early 1980s the need for asecond generation (2G) cellular system that would be totally digital This 2G system becameGSM, and a brief history of GSM has already been provided in Section 2.1 The Europeanshave a long view in cellular radio and in 1988 they launched their RACE 1043 project with

404

the aim of identifying the services and technologies for an advanced third generation (3G)system for deployment by the year 2000 [1, 2] Their 3G system soon became known asthe universal mobile telecommunications system (UMTS) [3–5] The concept was that their1G, 2G and 3G systems would be independent, and that their deployment would overlapsuch that the total access communications system (TACS), say, would slowly be replaced

by GSM, which in turn would be slowly phased out for UMTS However, the success ofGSM has been so great that evolutionary paths from 2G to 3G needed to be considered.Although the back-haul networks of GSM and UMTS have considerable commonality, theirradio interfaces are significantly different

There were initially great expectations for UMTS [5, 6] It would not only be cellular, but

it would embrace other types of networks from private mobile radio (PMR) (called specialmobile radio (SMR) in the United States), to wireless local area networks (WLANs), tomobile satellite systems (MSSs) The cardinal points were that it would operate globally,support high bit rate services and, most importantly, be service orientated While the Eu-ropeans referred to the global 3G network for the turn of the century as UMTS, most oftheir engineers working on UMTS expected that they would have to yield to internationalagreements from the ITU to modify UMTS, but that basically UMTS would be accepted asthe global standard

To explain this early expectation we need to point out that the ITU has been in the 3Ggame from the beginning [6] Paralleling the European Union (EU) RACE initiative, ITUformed task group TG8/1, originally under the auspices of CCIR This committee referred

to their 3G system as the future public land mobile telecommunications system (FPLMTS).Europeans were, of course, also members of TG8/1, and with commercial and political pres-sures a long way in the future, FPLMTS and UMTS seemed synonymous in terms of aimsand objectives The important difference between TG8/1 and the happenings in Europe,was that in Europe there was an actual research and development (R&D) 3G programme inprocess, while TG8/1 was more like a forum

The Americans did not launch concerted national R&D programmes, neither for 2G nor3G systems Their advanced mobile phone service (AMPS) 1G system did evolve into the2G IS-136, and became dual-mode with IS-95 The United States also introduced the iDENsystem with its ability to offer both cellular and dispatch services It then auctioned a largepart of its 3G spectrum for PCS licenses, and allowed GSM to enter the United States inthe form of PCS1900 This auctioning of the 3G spectrum meant that there were significantadvantages if existing 2G networks could evolve into 3G ones, preferably in a seamlessmanner

A big factor, not just in the United States, but in the world, was the advent of IS-95 [7–11]

It arrived late compared with GSM, and some engineers argued that it was a 2.5G system

It had to fight to be born because of the lack of spectrum, and the quasi-religious attitudes

of engineers towards methods of multiple access The meagre spectrum of 1.25 MHz at thetop of the AMPS band was just about adequate for cellular CDMA, which was just as wellbecause that was all there was available CDMA entered the cellular world with a host oftechnical problems, not made easier as the transceivers from day one had to be dual-modewith AMPS Its advocates were clear in that CDMA has a high spectral efficiency and iswell suited to the 3G multiservice environment The real significance of IS-95 is that it wonthe technical argument in that the UMTS and the Japanese Association of Radio Industriesand Businesses (ARIB) proposals have CDMA radio interfaces, albeit of wider bandwidthsystems as more bandwidth is available for 3G networks We therefore agree that IS-95 is a2.5G system and its evolution to 3G should be smooth This is not so for 2G TDMA systemswhich will need to migrate to 3G CDMA ones However, as we will show in Section 6.2,GSM with its TDMA is able to evolve closely to 3G without picking up the CDMA card.Nevertheless there is an evolutionary route from GSM Phase 2+ to UMTS as discussed inSection 6.2

The TG8/1 Committee discarded the unwieldy FPLMTS name for its 3G system, andreplaced it with international mobile telecommunications for the year 2000, or simply IMT-

2000 It then abandoned all hope of the difficult political objective of a single standard,and has instead opted for a family of standards Each member of the family had to beable to meet a minimum specification Sixteen proposals were accepted, ten for terrestrial3G networks, and six for MSSs The majority of the proposals advocated CDMA as themultiple access method A degree of harmonisation between the proposals ensued, and atthe time of writing the ITU has agreed that the IMT-2000 family will be composed of thefollowing five technologies

 IMT DS (Direct Sequence) This is widely known as UTRA FDD and W-CDMA,where UTRA stands for the UMTS Terrestrial Radio Access, and the ‘W’ in W-CDMA means wideband We will refer to this system here as UTRA FDD

 IMT MC (Multicarrier) This system is the 3G version of IS-95 (now calledcdmaOne), and is also known as cdma2000 We will use the term cdma2000 as this

is its widely used name

 IMT TC (Time Code) This is the UTRA TDD, namely the UTRA mode that usestime division duplexing

 IMT SC (Single Carrier) This is essentially a particular manifestation of GSM Phase2+, known as EDGE, standing for Enhanced Data Rates for GSM Evolution

 IMT FT (Frequency Time) This is the digitally enhanced cordless tions (DECT) system

telecommunica-In the authors’ opinion, the truly 3G systems are the IMT DS, IMT MC and IMT TCsystems.

6.1.2 IMT-2000 spectrum

The World Administration Radio Congress (WARC) in March 1992 assigned 200 MHz inthe 2G frequency band to IMT-2000 for world-wide use [3] The actual frequency bandsare 1885–2025 MHz and 2110–2200 MHz Unfortunately some parts of these bands arealready used for other services Figure 6.1 shows a diagram of the IMT-2000 spectrum, andthe current use of this spectrum in Europe, the United States, and Japan

The IMT-2000 spectrum may be partitioned into seven segments The frequency of eachsegment is shown in Table 6.1

Part of Segment 1 is currently used for DECT in Europe, and is also used for PHS, PCSand DECT in other parts of the world Segment 2 is used at present for PCS and PHS in theUnited States and Japan, respectively Segments 3 and 6 form 60 MHz frequency divisionduplex (FDD) bands Mobile satellite services (MSS) are in Segments 4 and 7, providing

30 MHz FDD bands Segment 4 supports the earth-to-space links; while segment 7 providesthe space-to-earth links The 1980–1990 MHz band in Segment 4 is currently used for PCS

in the United States Segments 1, 2 and 5 are unpaired and are suitable for time divisionduplex (TDD) operation Segment 5 may be used in the United States for earth-to-spaceMSS services

The GSM system was initially designed to carry speech, as well as low speed data Muchhas already been discussed regarding speech, so we will concentrate here on data The userdata rate over the radio interface using a single physical channel, i.e a single timeslot per

Table 6.1: IMT-2000 spectrum and its segments (MSS stands for mobile satellite services).

Figure 6.1: IMT-2000 spectrum and the current use of this spectrum in Europe, the United States

and Japan PCS stands for personal communication system, SAT for mobile satellite vices, DECT for digitally enhanced cordless telecommunications, and PHS for personalhandyphone system

ser-TDMA frame, was initially 9.6 kb/s The maximum user data rate available on a singlephysical channel has since been increased to 14.4 kb/s by reducing the power of the channelcoding on the full rate traffic channel by means of code symbol puncturing Apart fromincreasing the level of puncturing still further, the other ways to increase the user data ratebeyond 14.4 kb/s are either to allow an MS to access more than an one timeslot per TDMAframe or to use a higher level modulation scheme (e.g quadrature amplitude modulation,QAM) to increase the amount of information that can be transmitted within a single timeslot.Two new services have been introduced as part of GSM Phase 2+ which allow the userdata rate to be increased by permitting an MS to access more than one timeslot per TDMAframe These new services are the high speed circuit switched data (HSCSD) service andthe general packet radio service (GPRS) The HSCSD service allows an MS to be allocated

a number of timeslots per TDMA frame on a circuit-switched basis, i.e the MS has sive use of the allocated resources for the duration of a call [12] In contrast, GPRS usespacket-orientated connections on the radio interface (and within the network) whereby auser is assigned one, or a number of traffic channels only when a transfer of information

exclu-is required [13] The channel exclu-is relinquexclu-ished once the transmexclu-ission exclu-is completed In thefollowing sections we will describe these two services in more detail

The second approach to increasing the user data rate by employing a higher level lation scheme is currently being studied under the Enhanced Data Rates for GSM Evolution(EDGE) project [14] The basic principle behind EDGE is that the modulation scheme used

modu-on the GSM radio interface should be chosen modu-on the basis of the quality of the radio link Ahigher level modulation scheme is preferred when the link quality is ‘good’, but the systemreverts to a lower level modulation scheme when the link quality becomes ‘poor’ At thetime of writing it appears that EDGE will use the existing Gaussian minimum shift keying(GMSK) modulation scheme in poor quality channels and eight-level phase shift keying (8-PSK) in good quality channels EDGE will also include link adaption functions to allow the

MS and BS to assess the link quality and switch between the different types of modulation

as necessary

Once developed, the EDGE technology will enhance the range of services offered byGSM The initial version of the EDGE technology (Phase 1) will be used to enhance theGPRS and HSCSD services, leading to enhanced GPRS (EGPRS) and enhanced circuit-switched data (ECSD) In later releases of EDGE (Phase 2 and beyond) further serviceswill be introduced which utilise the different modulation schemes [14]

In addition to the developments described above, GSM Phase 2+ contains two other portant enhancements that have a significant impact on the technology from a radio point ofview In 1993 the European railways, in the form of the Union Internationale des Chemins

im-de Fer (UIC), chose the GSM technology as the basis of all their future mobile radio munication systems [15] This led to the introduction of a number of advanced speech callitems (ASCI) which provide the additional functionality required for railways and otherprivate mobile radio (PMR) environments The three key elements of ASCI are the voicebroadcast service (VBS), the voice group call service (VGCS) and the enhanced multi-levelprecedence and pre-emption (eMLPP) service The VBS will allow GSM users to broadcasttheir voice simultaneously to a number of other users in a chosen talk group A VBS callwill only occupy a single down-link channel in each cell within which the call is broadcastand all ‘listening’ MSs will monitor this same channel VBS calls are simplex in that thecall originator is the only person who can speak during the call Many applications requireany member of the talk group to become the talker and this functionality is supported in theVGCS In this case any member of the talk group may become the ‘talker’ and contentionresolution schemes are included to handle situations where more that one user tries to be-come the talker simultaneously These PMR systems also support a facility to ensure thatimportant calls are successfully completed even at the expense of less important calls TheeMLPP service allows calls to be prioritised and ensures that the most important calls arecompleted, regardless of the network loading

com-Another important Phase 2+ item from a radio perspective is the adaptive multi-rate(AMR) speech coder [16] The deployment of the GSM half-rate (HR) codec has been

somewhat limited because of concerns relating to the speech quality, but the enhanced rate (EFR) codec is more popular amongst GSM operators The basic concept behind theAMR technology is that the speech coding rate and the degree of channel coding should

full-be chosen according to the channel quality For example, in ‘good’ channels a lower ratespeech coder can be used in an HR traffic channel thereby increasing the system capacity.However, if the link quality is ‘poor’ FR then a traffic channel will be used and the level ofchannel coding increased At the time of writing the candidate AMR codecs have not yetbeen chosen

6.2.1 High speed circuit-switched data

The HSCSD service [12] is a natural extension of the circuit-switched data services thatwere supported in earlier versions of GSM No changes to the physical layer interfaces be-tween the different network elements are required for HSCSD At the higher layers the MSand the network support the additional functionality required to multiplex and demultiplex

a user’s data onto a number of traffic channels for transmission over both the Abis interfaceand the radio interface Additional functionality is also included at the radio resource man-agement level to handle the new situation where a number of different traffic channels areassociated with the same connection For example, when an HSCSD user is handed overbetween two cells, there must be a mechanism to ensure that sufficient traffic channels areavailable in the new cell before the handover occurs An HSCSD connection is, however,limited to a single 64 kb/s circuit on the A interface

On call set-up the MS provides information to the network which defines the nature of theHSCSD connection The multislot class of the MS is used by the network to determine themaximum number of timeslots that may be accessed by the MS, and the amount of time thatmust be allowed between timeslots, e.g for the purposes of neighbour cell measurements.This information is used to define the MS’s capabilities for both the HSCSD and GPRSservices The multislot classes are listed in Table 6.2 along with their associated parameters[17]

Multislot MSs can be either type 1 or type 2 and this information is shown in the hand column of Table 6.2 Type 2 MSs are required to be able to transmit and receivesimultaneously, whereas type 1 MSs are not The ‘Rx’ and ‘Tx’ columns give the maxi-mum number of receive and transmit timeslots that the MS may occupy per TDMA frame,respectively The ‘Sum’ column gives the total number of transmit and receive timeslotsthe MS may access per TDMA frame For example, for multislot class 12, ‘Sum’ is 5which means that the maximum number of transmit and receive slots cannot exceed 5 So

right-if we have 3 received slots, then we cannot have more than 2 transmit slots in one TDMA

frame The T taparameter represents the time required for the MS to make a neighbour cellmeasurement prior to an up-link transmission This parameter is not applicable to type 2

MSs because they are capable of making measurements and transmitting simultaneously.

When the MS is not required to make measurements on neighbouring cells, the T tb eter defines the minimum number of timeslots that must be allowed between the end of theprevious down-link timeslot and the next up-link timeslot, or the time between two con-

param-secutive down-link timeslots that are on different frequencies In other words, T tb is theamount of time the MS needs to prepare to transmit on the up-link after it has received or

transmitted on a different frequency The T raparameter is the number of timeslots required

by the MS to make a neighbour cell measurement prior to the reception of a down-link

burst, whereas the T rbparameter is the number of timeslots required between the previousup-link transmission and the next down-link reception, or the time between two consecutivedown-link receptions when the frequency is changed in between receptions In addition

to its multislot class, the MS provides a range of additional information on call set-up toallow the network to determine the most appropriate HSCSD configuration This infor-mation includes the fixed network user rate, i.e the data rate that the MS would like toachieve over the fixed network, the channel coding schemes supported by the MS, and themaximum number of traffic channels to be used during the connection This final param-eter allows the user to control the call cost by limiting the number of traffic channels thatwill be occupied The final multislot configuration is chosen by the network based on the

MS capabilities and the requirements imposed by the services, e.g whether neighbour cellmeasurements are required The HSCSD service can support both symmetric transmissions,i.e the same number of up-link and down-link timeslots, or asymmetric transmissions, i.e.more timeslots are allocated in one direction However, in the case of HSCSD connections,only down-link biased asymmetry is allowed and the up-link timeslot numbers must be asubset of the down-link timeslot numbers

6.2.2 The general packet radio service

Many services do not require a continuous bi-directional flow of user data across the radiointerface To illustrate this, consider the example of a user browsing the Worldwide Web(WWW) on her lap-top computer using a dial-up connection via a cellular network Once

a page of information has been downloaded, there will be a pause in the information flowbetween the MS and the network as the user reads the information and before more infor-mation is requested Using circuit-switched connections for ‘bursty’ services of this naturerepresents an inefficient use of the radio resources because a user will continue to occupy aradio channel for the duration of a call (or browsing session) even though this channel mayonly be utilised for a small fraction of the time Inefficiencies of this type can be overcome

by carrying these services using packet-orientated connections

The GSM system was initially designed to support only circuit-switched connections atthe radio interface level with user data rates of up to 9.6 kb/s However, the Phase 2+

Table 6.2: The MS multislot classes.

Multislot Maximum number of slots Minimum number of slots Type

1 if frequency hopping is used

0 if frequency hopping is not used

b =

(

1 if frequency hopping is used or there is a change from Rx to Tx

0 if frequency hopping is not used and there is no change from Rx to Tx

c =

(

1 if frequency hopping is used or there is a change from Tx to Rx

0 if frequency hopping is not used and there is no change from Tx to Rx

specifications now include provision for the support of a packet-orientated service known

as the general packet radio service (GPRS) [13, 18–20] GPRS attempts to optimise thenetwork and radio resources, and strict separation between the radio subsystem and thenetwork subsystem is maintained, although the network subsystem is compatible with theother GSM radio access procedures Consequently, the GSM MSC is unaffected Theallocation of a GPRS radio channel is flexible, ranging from one to eight radio interfacetimeslots in a TDMA frame Up-link and down-link timeslots are allocated separately Theradio interface resources are able to be shared dynamically between circuit switched andpacket services as a function of service load and operator preference Bit rates vary from

9 kb/s to more than 150 kb/s per user GPRS can interwork with IP and X.25 networks.Point-to-point and multipoint services are also supported, as well as short message services(SMS)

GPRS is able to accommodate both intermittent, bursty data transfers as well as largecontinuous data transmissions Reservation times are typically from 0.5 s to 1 s Three

MS modes are supported, each having a different arrangement with circuit switched GSMservices In this section we provide an overview of the GPRS technology and examine itsimpact on the GSM radio interface

Figure 6.2 is a block diagram showing the architecture of a GSM network that supportsGPRS, and the names that have been given to the interfaces that exist between the differentnetwork components GPRS services require two additional network components, the gate-way GPRS support node (GGSN), and the serving GPRS support node (SGSN) As its namesuggests, the GGSN acts as the gateway between external packet data networks (PDN), and

a GSM network that supports GPRS The GGSN contains sufficient information to routeincoming data packets to the SGSN that is serving a particular MS and it is connected toexternal networks via the Gi reference point (We note that this point of interconnection isreferred to a ‘reference point’ and not an ‘interface’ because no GPRS-specific information

is exchanged at this point.) The SGSN is connected by the Gn interface to GGSNs ing to its own public land mobile network (PLMN) and it is connected by the Gp interface

belong-to GGSNs belonging belong-to other PLMNs These two interfaces are very similar, but the Gpsupports additional security functions that are necessary for inter-PLMN communications.The GGSN may also interface directly with the home location register (HLR) over the Gcinterface, but this is not mandatory

A SGSN keeps track of the location information and the security information ated with the MSs that are within its service area A SGSN communicates with GGSNsand SGSNs in its own PLMN using the Gn interface and GGSNs in other PLMNs via the

associ-Gp interface Interfaces also exist between an SGSN and an MSC/VLR (Gs interface),

Gf Gn

SM-SC

Gi

PDN

Signalling Interface Signalling and Data Transfer Interface that support GPRS

Figure 6.2: The GPRS network architecture.

an HLR (Gr interface), an EIR (Gf interface) and a short message service gateway MSC(SMS-GMSC) and interworking MSC (SMS-IWMSC) (Gd interface) The SMS-GMSCand SMS-IWMSC allow the GSM short message service (SMS) to be carried on the GPRSchannels instead of by the SDCCH and the SACCH The GPRS support nodes (i.e theGGSNs and SGSNs) of a PLMN are interconnected using an Internet protocol (IP) basedbackbone network

This is a layered protocol structure enabling user information transfer and associated trol procedures such as flow control, error detection, error correction and error recovery.The GPRS transmission plane is shown in Figure 6.3 [18, 19] We observe in Figures 6.2and 6.3 the entities MS, BSS (namely BTS connected via the Abis to the BSC), SGSN andthe GGSN; together with the interfaces Um, Gb, Gn and Gi From the transmission plane

con-we observe that at the highest level the application is leaving the network via Gi The GPRStunnelling protocol (GTP) tunnels user data and signalling between GPRS support nodes,SGSN and GGSN, in the backbone network The transport control protocol (TCP) carriesthe GPRS tunnelling protocol (GTP) data units (PDUs) in the GPRS backbone network for

protocols that need a reliable data link, such as X.25 The user datagram protocol (UDP)conveys GTP PDUs for protocols, such as Internet protocol (IP) that do not require reliablelinks The TCP provides flow control and protection against lost and corrupted GTP PDUs.The GPRS backbone network protocol is IP, and is used for routing user data and signalling.The subnetwork dependent convergence protocol (SNDCP) maps network-level character-istics onto the lower layer network The provision of a highly reliable ciphered logical link

is achieved by the logical link control (LLC) The LLC is independent of the radio interfaceprotocols

In the BSS the relay function relays LLC PDUs between the radio interface Um and the

Gb interface In the SGSN the relay function relays packet data protocol (PDP) PDUsbetween the Gb and the Gn interfaces The BSS GPRS protocol (BSSGP) conveys routingand quality-of-service (QoS) information between a BSS and an SGSN It does not performerror control The network service (NS) layer transports BSSGP PDUs The radio linkcontrol (RLC) provides a radio dependent reliable link, while the medium access control(MAC) controls the access signallings, both request and grant, for the radio channel; andthe mapping of the LLC frames onto the GSM physical channel GSM RF is the GSMphysical layer

registration, authentication and authorisation, admission control, charging and so on, thatallow a user to use the services and facilities the network provides A user may make anaccess attempt, or the user may be paged The fixed network interface, Gi, may support, atthe discretion of the PLMN operator, multiple access protocols to external networks such

as X.25 and IP

whereby an MS can access the network without being authenticated and without air-interfaceencryption being required There is also no requirement for the MS to supply its IMSI orIMEI, although there is provision for the network to request these One example of an ap-plication that could make use of this facility is automatic road-tolling, whereby a road usercould use a pre-paid card, inserted into a GSM terminal, automatically to make payment asshe approaches a toll booth

and if required, relay nodes, and finally a destination node Routing is the transmission

of messages within and between PLMNs A node forwards data received to the next nodeusing the relay function The routing function determines the network GPRS support node

L1 bis

GbBSS

Figure 6.3: GPRS transmission plane [18–20].

(GSN) where the message is sent using the destination address in the message The routingfunction also selects the transmission path to the next GSN along the route Address transla-tion is needed to convert one address to another when packets are routed between PLMNs.Encapsulation is the addition of address and control information to the PDU for routingwithin and between PLMNs Encapsulation, and its reverse, are performed between theGGSN nodes of PLMNs, and between the SGSN and an MS The tunnelling function (seethe transmission plane) is the transfer of encapsulated PDUs within and between PLMNsfrom where they are encapsulated to where they are decapsulated There is a compressionfunction that removes as much overhead information as possible, prior to radio transmis-sion The ciphering function provides confidentiality of a user’s data, while the domainname server function is a standard IP function that resolves any name for GSNs and other

GPRS nodes within the GPRS backbone networks.

processes in the mobile radio system that are associated with tracking the movements ofsubscribers as they travel around a network The non-GPRS GSM system uses locationareas and location update procedures to ensure that it always knows the whereabouts of itsMSs The GPRS uses a similar approach, but instead of using the existing location areas(LAs) it uses routing areas (RAs) If an MS detects that it has entered a new RA it willsubmit a routing area update request message to the network, which if successful, will beacknowledged with a routing area update accept message The decision as to whether aGPRS-equipped MS should perform the RA update procedure as it moves between RAswill depend on the GPRS state of the MS

A GPRS-equipped MS may be in any one of three different states for the purposes of

mobility management In the idle state the MS is ‘not reachable’ as far as the GPRS is

concerned If an MS is in the idle state, the network does not hold any information regardingthe location of the MS, and hence the MS cannot be paged It also means that the MS doesnot need to perform any RA updates as it moves around the network MSs in the idle state

cannot access the packet data services without first performing a procedure known as GPRS

attach where the MS announces its identity.

In the standby state the user is attached to the GPRS mobility management (MM) The

MS performs RA and cell selection, and may receive point-to-multipoint multicast M) and point-to-point multipoint group call (PTM-G) data The SGSN may send data or

(PTM-signalling information to an MS The MM state in the MS changes to ready when an MS

re-sponds to a page, as does the state in the SGSN when the response from the MS is received.The MM state in the MS goes to ready when data or signalling information is sent by the

MS to the network, hence the MM in the SGSN also changes to the ready state Either the

MS or the SGSN may initiate the GPRS detach procedure to move to the idle state

In the ready state, the MS informs the network of the selected cell by means of an identifierincluded in the BSSGP header of the data packet from the MS The MS is able to send andreceive PDP PDUs The network initiates no pages for an MS in the ready state, but pagesfor other services may be executed via the SGSN The ready state is supervised by a timer,and when the timer expires the MM moves from ready to standby To move from ready toidle states, the MS initiates the GPRS detach procedure

main-tenance and release procedures between an MS and the PLMN over the radio interface.These functions involve the co-ordination of link state information and the supervision ofdata activity over the logical link

Radio resource management functions Allocation and maintenance of radio cation channels are provided by these functions The GSM radio resources are dynamicallyshared between the circuit mode and GPRS The GPRS radio resource management is con-cerned with the allocation and release of timeslots for a GPRS channel; monitoring GPRSchannel utilisation; congestion control; and the distribution of GPRS channel configurationinformation that is broadcast on the common control channels.

GPRS

6.2.2.4Low-level functions

The radio interface functions for GPRS are the medium access control (MAC) and radiolink control (RLC) that operate above the physical layer The MAC function arbitrates be-tween MSs attempting to transmit at the same time It is therefore concerned with collisionavoidance, detection, and recovery following a collision The MAC function may let a sin-gle MS use several physical channels simultaneously The multiplexing of data and controlsignalling on both links is affected by the MAC function, as well as by priority scheduling.The GPRS RLC function supports the transfer of logical link control layer PDUs (LLC-PDU) between the LLC and MAC entities, the segmentation and reassembly of LLC-PDUsinto RLC data blocks, and backward error correction for the retransmission of uncorrectablecode words

physical channels, they have different logical channels The physical channel dedicated topacket data is called a packet data channel (PDCH) The logical channels for common con-trol signalling for packet data are carried by the packet common control channel (PCCCH),and there is also the packet random access channel (PRACH), an up-link-only channel used

by MSs to initiate data or signalling packet transmission The other PCCCHs are all link ones There is the packet paging channel (PPCH) that uses paging groups of MSs

down-to enable discontinuous reception (DRX) down-to be used PPCH can be used for both circuitswitched and packet services The packet access grant channel (PAGCH) identifies the re-source assignment to be used by an MS prior to packet transfer, while the packet notificationchannel (PNCH) provides notification to a group of MSs that a PTM-M packet transfer isimminent The packet broadcast control channel (PBCCH) informs the MSs of packet dataspecific information This information may also be transmitted on the BCCH if a PBCCHhas not been allocated

The packet data traffic channel (PDTCH) is allocated for data transmissions It is porarily allocated to an MS (or a group of MSs in the PTM-M case) An MS may be

tem-assigned multiple PDTCHs.

There are also packet dedicated control channels The packet associated control channel(PACCH) conveys signalling information to a specific MS The type of information includesacknowledgements, power control, resource assignment and reassignment The PACCH andPDTCH share the resources allocated to an MS An MS that is transferring a packet can bepaged for circuit switched services on the PACCH Another dedicated control channel is thepacket timing advance control channel, up-link (PTCCH/U) that conveys in random accessbursts the information to allow the network to deduce the necessary timing advance for the

MS packet transmissions There is also the packet timing advance control channel, link (PTCCH/D) that transmits timing advance information updates to MSs One PTCCH/D

down-is paired with several PTCCH/Us

A number of PDCHs can share the same physical channel They are mapped dynamicallyonto a 52-multiframe consisting of 12 blocks of four consecutive frames, two idle framesand two frames for the PTCCH, as shown in Figure 6.4 The 52-multiframe has a duration

of 240 ms, namely two 26-multiframes of the GSM TCH The first of the 12 blocks in the52-multiframe is B0 and its PDCH contains PBCCH On any PDCH with PCCH, up to

12 blocks can be used for PAGCH, PNCH, PDTCH or PACCH on the down-link On theup-link PDCH that contains PCCCH, the blocks in the multiframe can carry the PRACH,PDTCH or PACCH The mapping of channels onto multiframes is controlled by parametersbroadcast on the PBCCH

There are three classes of MS as far as the GPRS is concerned and these are based onthe ability of the MS to support the simultaneous use of packet-based and circuit-switched

services The Class A MSs are able to support the simultaneous transfer of both

packet-based and circuit-switched traffic using different timeslots within the GSM TDMA frame

structure The Class B MSs can be simultaneously ‘attached’ to both the circuit switched

and packet-based services, e.g they can receive pages for either service; however, theycannot transfer packet-based traffic and circuit-switched traffic at the same time If, forexample, a Class B MS is engaged in a packet-based data transfer and a circuit-switchedconnection is established, the transfer of packet data will be suspended for the duration of

Figure 6.4: Multiframe structure for PDCH.

the circuit-switched connection However, in this example there is no requirement for the

GPRS to be deactivated for the particular MS Class C MSs do not support any simultaneous

use of circuit-switched and packet-based services A Class C MS can be viewed as operating

in two distinct modes, the GPRS mode and the circuit-switched mode, and it can only be

in one mode at a particular time For example, if the class C MS is in GPRS mode (i.e it

is engaged in a packet-based data transfer) and a circuit-switched call arrives at the PLMNfor this particular MS, then the MS will be considered ‘not reachable’ as far as the circuit-switched call is concerned It should be noted that the GPRS class of an MS and its multislotclass (see Table 6.2) are separate parameters and there is no direct correlation between thetwo

There are four different quality-of-service (QoS) profiles supported in GPRS and these aredependent on the type of service For example, an email transaction can tolerate greater de-lays than, say, an interactive video service, and the QoS profile will be chosen appropriately

in each case The MS and the network will agree on a particular QoS profile during theinitial service negotiation stages and, as far as possible, the network will attempt to deliverthis QoS to the MS The QoS service profile is made up of a number of factors, includingthe delay class (i.e the average packet transmission delay), the precedence class (i.e thepriority value attached to the packets in the event of packet erasure being required as a re-sult of network congestion), the reliability class (i.e the probability of errors in the receiveddata packet), and the peak and average throughput class (i.e the peak and average rates atwhich data are transferred through the network, respectively)

6.2.3 The enhanced data rates for GSM evolution (EDGE)

The driving force behind EDGE is to improve the data rates of GSM by means of ing the modulation methods [14, 21]; specifically, to increase the data rate transmissionper radio timeslot compared with GMSK modulation Different types of enhanced modu-lation methods have been considered starting with quaternary offset quadrature amplitudemodulation (Q-O-QAM) and binary offset quadrature amplitude modulation (B-O-QAM),and ending with 8-level phase shift keying (8-PSK) Although the initial drivers were toincrease the user bit rates and thereby increase the range of services, EDGE has been gilded

enhanc-as a 3G system and is now a member of the IMT-2000 family EDGE’s new heady role henhanc-as

a lot to do with the evolutionary strategy of IS-136, the US TDMA system that is itself anevolution of the former analogue system, AMPS The Universe Wireless Communications(UWC) Consortium advocated a family of mutually compatible TDMA standards known asUWC-136 that would be developed from the second generation IS-136 The concept was

that voice services would be provided by the existing IS-136 network and by a new variantcalled 136+ which would employ enhanced modulation methods, such asπ=4 differentialquadrature phase shift keying (π=4 DQPSK), coherent QPSK, and coherent 8-PSK offering

a maximum user rate of 43.2 kb/s Higher rate services would be provided by 136 HS,which supports user rates up to 521 kb/s for pedestrian mobiles and for vehicular mobiles atspeeds up to 100 km/h For speeds in the range of 100–500 km/h, rates of 182 kb/s would

be accommodated In indoor environments, the maximum bit rate would be 4.7 Mb/s For

136 HS the carrier spacing is 200 kHz (as in GSM) for outdoor/vehicular environments, and

1600 kHz for offices

The 136 HS (outdoor/vehicular) with its 200 kHz carrier spacing, eight slots per frame,FDD, with multilevel modulation was part of the UWC-136 IMT EDGE proposal Theevolution of GSM to EDGE and IS-136 to EDGE is now undertaken jointly by ETSI andthe UWC Consortium Consequently, EDGE will be compatible with both GSM and IS-

136 The plan [14] is to deploy GPRS, then enhanced GPRS (EDPRS) and enhanced circuitswitch data (ECSD) Then the high level of modulation will be deployed to realise 3GEDGE services

will be able to use 8-PSK which has three bits per symbol instead of the one-bit type symbol

of GMSK Since the symbol rate is 271 ksymbols/s then the gross bit rates per slot (includesoverhead) is 22.8 and 69.2 kb/s for GMSK and 8-PSK, respectively The pulse shape for8-PSK is such that the 8-PSK spectrum fits within the GMSK spectrum mask The normalburst format for EDGE is the same as for GSM, except the two sets of 58 symbols now havethree bits per symbol

switched modes Indeed, EDGE is more like the grand evolutionary plan of GSM that cludes both GPRS and HSCSD The enhanced GPRS (EGPRS) differs from GPRS becausewith multilevel modulation the channel coding must be improved because it is more vul-nerable to interference and noise Accordingly a link adaption scheme regularly estimateslink quality and selects GMSK or 8-PSK and the appropriate channel coding to provide thehighest user bit rate At the time of writing various schemes are being considered for stan-dardisation The coding rate is determined by the amount of puncturing The rate per timeslot for GMSK is 11.2, 14.5, 16.7 and 22.8 kb/s for code rates of 0.49, 0.64, 0.73, and 1.0,respectively For 8-PSK the rate per time slot is 22.8 34.3, 41.25, 51.6, 57.35 and 69.2 kb/sfor code rates of 0.33, 0.50, 0.60, 0.75, 0.83 and 1.0, respectively

in-The enhanced circuit switched (ECSD) mode has the data interleaved over 22 TDMAframes For GMSK modulation the rate per time slot is 3.6, 6, 12 and 14.5 kb/s for a code

rate of 0.16, 0.26, 0.53 and 0.64, respectively; while for 8-PSK the bit rates have the highervalues of 14.5, 29, 32 and 38.8 kb/s for code rates of 0.42, 0.46, and 0.56, respectively.For EGPRS when QoS issues are addressed where different factors, such as priority ofpacket transmissions, packet delay, packet throughput, maximum and minimum bit ratesthat must be handled, and so on, need to be considered, then the maximum bearer rateshould be able to accommodate data rates of 384 kb/s for MS speeds up to 100 km/h and

144 kb/s for an MS travelling at 250 km/h These rates require a user to make use of multipleslots per frame if these high bit rates are to be achieved Fewer slots are required if 8-PSK

is used Similar comments can be made regarding bearer rates for ECSD For low bit errorrate transmissions the maximum user bit rate is 57.6 kb/s

Because of different user requirements there is one class of MSs where 8-PSK is used

in the down-link and only GMSK in the up-link This provides higher bit rate down-linktransmissions than the up-link ones while at the same time decreasing the complexity of the

MS Another class of MS will support 8-PSK transmissions on both links

As with all the 3G standards, we await with interest to see how well they will performwhen they are deployed in operational networks

6.3 The Universal Mobile Telecommunication System

The universal mobile telecommunication system (UMTS) is, at the time of writing, ing shaped within the Third Generation Partnership Project (3GPP) [22] The participantshave come together for the specific task of specifying a 3G system based on an evolvedGSM core network and the UTRA FDD and TDD radio interfaces The 3GPP is composed

be-of organisational partners, market representation partners and observers The organisationpartners, i.e standards organisations, are: ARIB (Japan), CWTS (China), ETSI (Europe),

TI (USA), TTA (Korea) and TTC (Japan) The market representation partners are: GlobalMobile Suppliers Association (GSA), the GSM Association, the UMTS Forum, the Univer-sal Wireless Communications Consortium (UWCC), and the IPv6 Forum The observersare TIA (USA) and TSACC (Canada) The 3GPP activity is overseen by a project coordina-tion group (PCG) The specifications are developed by four technical specification groups(TSGs) responsible for the core network, the radio access networks, services and systemaspects, and terminals Each TSG has a number of working groups There will be a roll-out

of the specifications; the initial release of the specifications is Phase 1 Release 99 Newcapabilities and services will be introduced according to annual specification releases.The UMTS terminology introduces a number of new terms, and re-names some familiarones Many of these new terms will be defined as they appear in the text, but to assist thereader give a list of UMTS abbreviations in Table 6.3 We also draw the reader’s attention

to some familiar GSM terms that are different in UMTS These are presented in Table 6.4

Table 6.3: UMTS abbreviations.

ACLR Adjacent channel leakage power ratio

AI Acquisition indicator

AICH Acquisition indication channel

BCH Broadcast control channel

CCPCH Common control physical channels

CPCH Common packet channel

CPICH Common pilot channel

DPCCH Dedicated physical control channel

DPDCH Dedicated physical data channel

DCH Dedicated channel

FACH Forward access channel

FBI Feedback information

FDD Frequency division duplex

GGSN Gateway GPRS support node

IuCS Interface between an RNC and an MSC

IuPS Interface between an RNC or BSC and an SGSN

Iur Interface between RNCs

MAC Medium access control

MSC Mobile switching centre

MUD Multiuser detection

Node B Base station transceiver

OVSF Orthogonal variable spreading factor

P-CCPCH Primary common physical channel

PCH Paging channel

PCPCH Physical common packet channel

P-CPICH Primary CPICH

PI Paging indicator

PICH Pilot channel

PRACH Physical random access channel

PSC Primary synchronisation code

QPSK Quadrature phase shift keying

RACH Random access channel

RNC Radio network controller (like BSC in GSM)

RNS Radio network subsystem

S-CCPCH Secondary common control physical channel

S-CPICH Secondary CPICH

SCH Synchronisation channel

SF Spreading factor

SGSN Serving GPRS (generalised packet rate service)

SSC Secondary synchronisation code

TFCI Transport format combination indicator

TPC Transmit power control

UARFCN UTRA absolute radio frequency channel number

UMTS Universal mobile telecommunication system

USIM Universal subscriber identity module

UTRA UMTS terrestrial radio interface

The UMTS network architecture block diagram is displayed in Figure 6.5 The core work is encased by a dotted line The mobile switching centre (MSC) and gateway MSC(GMSC) are for circuit-switched GSM networks Because GSM Phase 2+ will also ac-commodate GPRS, and therefore handle packet data, there is a serving GPRS support node(SGSN) and a gateway GPRS support node (GGSN) The other core network elements to dowith authentication, home and visitor location registers and equipment identity registers areessential to support both circuit-switched and packet data networks Thus, the core network

net-is architecturally a GSM Phase 2+ core network that net-is powered up so that it can also handlethe higher volume, higher bit rate, UMTS traffic

Shown below the core network in Figure 6.5 are two GSM base station subsystems (BSSs)and two UMTS radio network subsystems (RNSs) The A-interface is between a base sta-tion controller (BSC) and a mobile switching centre (MSC), and there is an IuPS between

a BSC and SGSN, where the subscript uPS signifies a packet switch interface The Abisinterface between a BTS and a BSC is also shown

The UMTS network uses the same core network as GSM, and has interfaces betweenthe RNC and MSC, SGSN and RNC of IuCS, IuPSand Iur, respectively The subscript uCS

Table 6.4: GSM and UMTS terminologies of some key entities.

Mobile station (MS) User equipment (UE)

Base station transceiver (BTS) Node B

Base station controller (BSC) Radio network controller (RNC)

Base station subsystem (BSS) Radio network subsystem (RNS)

Subscriber identity module (SINM) Universal subscriber identity

module (USIM)

AuC = Authentication Center

HLR = Home Location Register

VLR = Visitor Location Register

EIR = Equipment Identity Register

= Traffic and Signalling

MSC = Mobile Switching Centre SGSN = Serving GPRS Support Node GGSN = Gateway GPRS Support Node

= Signalling Only

Modified from 3GPP TS 23.002 Version 3.1.0

Core Network

To external networks To external networks

represents circuit-switched The equivalent of the Abisin UMTS is Iubis.

This figure is important because it illustrates the evolutionary path from GSM Phase2+ toUMTS We see the different radio interfaces of GSM and UMTS plugging into a commonbackbone network

The physical channels in UMTS transfer information across the radio interface In theFDD version of UMTS, a physical channel is defined by its code and carrier frequency,while in TDD it is in terms of its code, carrier frequency and timeslot The physical layer(layer 1) of the protocol stack supports transport channels to the medium access control(MAC) in layer 2 The MAC layer offers logical layers, e.g radio link control, to the higherlayers A logical channel is characterised by the type of information transferred, e.g it may

be handling control information of data traffic

Figure 6.6 is a block diagram of a UMTS transmitter at the physical layer Transportchannel data from layer 2 and above are arranged in blocks depending on the type of data.These blocks are cyclically redundancy coded (CRC) to facilitate error detection at thereceiver The data are segmented into blocks and channel coding ensues The coding may

be convolutional or turbo Sometimes channel coding is not used Data are then interleaved

to decrease the memory of the radio channel and thereby render the channel more like The interleaved data are then segmented into frames compatible with the requirements

Gaussian-of the UTRA interface Rate matching is performed next This uses code-puncturing anddata repetition, where appropriate, so that after transport channel multiplexing the data rate

is matched to the channel rate of the dedicated physical channels A second stage of bitinterleaving is executed, and the data are then mapped to the radio interface frame structure.Suffice to say at this point is that there are different types of physical channels, namelypilot channels that provide a demodulation reference for other channels; synchronisationchannels that provide synchronisation to all UEs within a cell; common channels that carryinformation to and from any user equipment (UE); and dedicated channels that carry infor-mation to and from specific UEs

The physical layer procedures include cell search for the initial synchronisation of a UEwith a nearby cell; cell reselection which involves a UE changing cells while not engaged in

a call; access procedure that allows a UE to initially access a cell; power control to ensurethat a UE and a BS transmit at optimum power levels; and handover, the mechanism thatswitches a serving cell to another cell during a call

6.3.1 The UTRA FDD mode

The UMTS terrestrial radio interface (UTRA) frequency duplex (FDD) mode is the CDMA radio interface of the UMTS, and is designated by the ITU as IMT DS Referring toTable 6.1, the UTRA FDD mode uses segment 3 for up-link transmission, and segment 6 fordown-link transmission, i.e from node B to UE communications These two segments are

W-Transport Channel Data From Higher Layers

CRC Attachment

Block Concatenation/

Segmentation

Channel Coding

1 st Interleaving

Radio Frame Segmentation

Rate Matching

Traffic Channel Multiplexing

Physical Channel Segmentation

2 nd Interleaving

Physical Channel Mapping

To radio interface

Convolutional Coding Turbo Coding

No Coding Other

Channels

Figure 6.6: Block diagram of a UMTS transmitter at the physical layer.

called band A (There is a band B for PCS services in the United States, where 1850–1910and 1930–1990 MHz are for the up-link and down-link transmission, respectively)

For the UTRA FDD the duplex spacing, i.e the frequency separation between pairedchannels, is in the range 134.8 MHz to 245.2 MHz, and all UE must support a duplexspacing of 190 MHz The nominal spacing between radio carriers is 5 MHz, with a chan-nel raster of 0.2 MHz This means that the carrier separation may be adjusted in steps of0.2 MHz, e.g the carrier spacing may be 4.8 MHz

The carrier frequency is defined by the UTRA absolute radio frequency channel number(UARFCN) This number is defined over a frequency band from 0 to 3.7 GHz, and it is thetransmission frequency multiplexed by five Consequently, the UARFCN, which we denote

as N u and N d for the up-link and down-link, respectively, will always be an integer because

of the raster frequency of 0.2 MHz Note that the radio channels in the UTRA FDD are notnecessarily paired as they are in GSM

spec-trum must conform to, i.e be within, a specspec-trum mask [23] As an example, the mask for a

BS with a maximum output power of43 dBm is shown in Figure 6.7 The abscissa is thefrequency offset,4f , from the carrier The carrier is at 4f =0, not shown in the figure,and the adjacent carrier would be positioned at4f =5 MHz The lightly shaded part ofthe figure corresponds to the left-hand ordinate, which is the power measured by a spectral

analyser using a bandwidth of 30 KHz The maximum output power as a function of4f

between 2.7 and 3.5 MHz must be less than 14 15(4f 2:7)dBm The darker shadedarea of the figure is associated with the right-hand ordinate, which is the power measured in

a 1 MHz bandwidth The value of4f maxis 12.5 MHz, or the edge of the allotted frequencyband for UMTS, whichever is the greater

The adjacent channel leakage power ratio (ACLR) is the ratio of the transmitted power

to the power measured in an adjacent band following a receiver filter Both the transmittedand received power are measured at the output of a raised cosine filter with a 0.22 roll-offfactor and a noise power bandwidth equal to the chip rate For an adjacent channel offset of

5 MHz i.e 4f=5 MHz in Figure 6.7, the ACLR limit for the BS is 45 dB less than thetransmitted signal power, while for a UE it is 33 dB down on the transmitted power of theadjacent carrier, or if the transmitted power of the adjacent carrier is 50 dBm, whichever

is the higher The corresponding figures for4f=10 MHz are 50 dB for the BS, and 43 dB

or 50 dBm, whichever is the greater, for the MS

and that for UTRA FDD they are defined by a specific carrier frequency and code There

are two basic types of physical channels, called dedicated channels and common channels.

The former are used by UEs for the duration of a call, while the latter carry information toall UEs within a cell and are used by the UEs to access the network

There are two types of dedicated physical channels The dedicated physical control nel (DPCCH) carries physical layer (i.e layer 1) control information and the dedicatedphysical data channel (DPDCH) transports the user traffic, as well as control informationfrom layer 2 and from higher layers

chan-The DPCCH contains pilot symbols, transmit power control (TPC) symbols, and a port format combination indicator (TFCI) The pilot symbols enable the receiver to estimatethe impulse response of the radio channel and to perform coherent detection They are alsonecessary when adaptive antennae are used that have narrow beams The pilot symbolsconstitute a pilot word having a duration of 0.667 ms

trans-The TPC commands the fast closed-loop power control, and are used on both the up-linkand down-link TPC symbols are included in every transmitted packet, and they convey

a binary instruction, namely to increase or decrease the transmitted power by a specificamount

The TFCI informs the receiver of the instantaneous parameters of the different transportchannels; that is, it tells the receiver of the data rates currently in use The TFCI alsocontains feedback information (FBI) on the up-link which is used to provide a feedbackloop for transmit diversity and selection diversity

The frame structure, multiplexing arrangement and spreading schemes are different for

Frequency offset, ∆f, from carrier (MHz)

Figure 6.7: Transmitted spectrum mask for a UMTS BS.

the up-link and the down-link The DPCCH and DPDCH are time multiplexed i.e theDPCCH and DPDCH are multiplexed within a slot prior to spreading and transmission

at a BS By contrast the up-link transmissions from a UE have the DPDCH and DPCCHtransmitted in parallel on the quadrature components of the dual-channel QPSK modulator

The down-link frame structure is shown in Figure 6.8 The duration of a frame is 10 ms,and contains 15 slots Each slot has a duration of 0.667 ms and contains 2560 chips as thechip rate is 3.838 Mchips/s The actual data per slot is 102k , k=0;1; : ;7, namely 10 to

1280 bits, where k is a system parameter The spreading factor is

by the next DPCCH having N T PC , followed by a second DPDCH having N data2bits of data

2, and finally the last DPCCH with N pilot bits of pilot signal This means that when twoadjacent frames are considered, the pilot bits are placed between data 2 bits and data 1 bits

We also note that, unlike GSM where sounding is executed every 4.6 ms, UTRA sounds theradio channel every 0.667 ms The greater sounding rate of UTRA is necessary because ofthe higher bit rate services supported by UTRA compared with those in GSM

Frame 0

Slot 2

Frame 2

Slot i

Frame i

Slot 14

Frame 71

Frame 0

Slot 2

Frame 2

Slot i

Frame i

Slot 14

Frame 71

Figure 6.8: Down-link slot and frame structure for dedicated channels.

There are 17 different down-link slot formats that relate to the channel bit rate For

exam-ple, if the bit rate is 15 kb/s then N data1=2;N data2=2;N T FCI=0;N T PC=2;and N pilot =4,

totalling 10 bits in a slot, and corresponding to k=0;SF=512 We see that in this situationthere are only four bits out of 10 bits carrying data, corresponding to an overhead of 60%!

For a channel bit rate of 1920 kb/s (slot format 16), N data1=240;N data2=1008;N T FCI=

8;N T PC=8 and N pilot =16, giving a total of 1280 bits per slot and SF =4 The overheadnow is only 2.5%

All the down-link slots contain pilot symbols, and the pilot bit patterns are defined for

N pilot values of 2, 4, 8 and 16 bits For a particular N pilot, the logical values of the bitschange from slot to slot creating a pilot pattern that repeats for each 15-slot frame By thisprocedure the pilot symbols affect both frame synchronisation as well as channel estimation

The TPC field has N T PC=2;4 or 8 bits, and these N T PCbits are either all logical 1s or alllogical 0s, depending on whether the power control command is to have the UE increase ordecrease its transmitter power by the step size, respectively

The TFCI provides information about the data rates currently being used on the datachannels, e.g in the case of multiple simultaneous services The TFCI may be omitted,e.g in the presence of fixed rate services The UE will know the down-link slot format,

and hence that N T FCI =0 It will also know the value of the SF, the pilot pattern, and

N data1;N data2 , and N T PC (but not their logical values) The UE must apply blind detection

in which it must try different decoding strategies that are likely to be the inverse of those

used in conditioning the data The decoding strategies that provide the lowest bit error rate(BER) is deemed to be the correct one.

We have seen that the DPCCH carries the TFCI, TPC and pilot bits on a time divisionbasis within each slot, while the DPDCH carries data 1 and data 2 also on in a time divisionmode within a slot Further more, both the DPCCH and DPDCH are multiplexed together.The DPCCH and DPDCH signals are converted from a serial stream to two parallel streams,one called the inphase component and the other the quadrature component At this pointalthough they are designated as inphase and quadrature components, they are, in fact, of thesame phase Both bit streams are then spread to 3.84 Mchips/s using the same code, known

as the channelisation code, C ch;SF;n , and SF=2n Note that if a bit from the converter is alogical 0, then the code replaces the bit, and if the bit is a logical 1, then the code is inverted

The SF must be chosen so that the chip rate is always 3.84 Mchips/s As an example, if the combined DPDCH/DPCCH is 960 kb/s, then the I and Q branches will operate at 480 kb/s, and the SF is 3.84 Mchips/s divided by 480 kb/s, namely 8.

If the required data rate exceeds the capacity of a single DPDCH channel, then additional

DPDCHs can be added Suppose N DPDCHs are required, then we still need only one

DPCCH, but each DPDCH requires its own channelisation code The arrangement is shown

in Figure 6.9 [24], where the DPCCH is multiplexed with DPDCH1 The absence of anycontrol data on DPDCHi , i=2;3; : ;N;means that transmissions are suspended duringthose portions of the slots when DPCCH transmission would normally occur The down-link slot format for these multicode transmissions is displayed in Figure 6.10 [25]

The so-called inphase and quadrature spread signals from each set of multipliers in

Fig-ure 6.9 are applied to adders to create the so-called inphase signal I and quadratFig-ure signal

degrees so that the I and Q signals are now really in quadrature, and on adding these signals we have I+jQ.

This resulting signal is scrambled by a code, C scramb, but before discussing this code wewill describe the channelisation codes

(OVSF) codes that identify the down-link channel They are basically Walsh codes of ferent lengths that are able to preserve orthogonality between channels even when they areoperating at different data rates The OVSF codes are arranged in a tree structure for codeallocation purposes The OVSF code tree is shown in Figure 6.11 For a spreading factor of

dif-SF=1 there is a channelisation code C ch; 1 ; 0 = (1), i.e a word of one bit that is a logical 1

For SF=2, there are two codes C ch; 1 ; 0 = (1;1)and C ch; 2 ; 1 = (1; 1) Doubling the SF to 4 gives four codes: C ch; 4 ; 0 = (1;1;1;1);C ch; 4 ; 1 = (1;1; 1; 1);C ch; 4 ; 2 = (1; 1;1; 1); and

C ch; 4 ; 3 = (1; 1; 1;1) We observe that C ch; 4 ; 0is C ch; 2 ; 0followed by C ch; 2 ; 0, i.e the C ch; 2 ; 0

and its repeat, whereas C ch41 is C ch20followed by its C ch20but with its bits inverted In

Figure 6.9: Down-link spreading arrangement for multicode transmission.

One Slot = 2560 chips

Channel 2

Channel N

Modified from 3GPP TS 25.211 V3.0.0

Figure 6.10: Down-link slot format for multicode transmission.

an illustration in support of this criterion, suppose a cell is using C8; 4 If it then decides

to use the longer code C16 ; 8the cross-correlation between two concatenated C8 ; 4codes and

the C16; 8 code is non-zero and hence there will be interference if these codes are used By

selecting C8; 3and C16; 8 on different branches the codes are uncorrelated This situation isillustrated in Figure 6.13

Observe that UMTS has codes of different lengths simultaneously in use in a cell, pending on the services being handled Good code allocation is essential if a large number

de-of codes are to be available For example, if all the short codes are used, then no long codesare available, thus restricting capacity Care must also be exercised in selecting codes toprevent the crest factor, i.e the ratio of the peak-to-mean transmitted power, from becom-ing too high The crest factor is allowed to be up to 18 dB We recall that in the presence

of multiple users the amplitude values of the I and Q signals are dependent upon the sum

of the user codes If during any chip interval the codes sum to a large number, then thepeak power becomes large This makes it necessary for the transmitter equipment, i.e thepower amplifier and filters, to be specified for these surges in power A low crest factoreases equipment design and cost

of fading, non-zero cross-correlation functions This renders them unsuitable for ple access codes Since the channelisation codes are Walsh codes of differing lengths itfollows that scrambling these codes by another code having good autocorrelation and cross-correlation codes is essential This arrangement is used in cdmaOne The scrambling codeemployed in UTRA FDD is a 38 400-chip segment of a 218 1 length Gold code Thescrambling codes have inphase and quadrature components, and there are a total of 8192codes For the 3.84 Mchips/s transmitted chip rate, the 38 400 chip code lasts 10 ms, i.e theduration of the scrambling codes lasts for one 15-slot frame

multi-The 8192 codes are divided into 512 sets, with each set having 16 codes multi-These 16 codesare composed of a primary scrambling code, and 15 secondary scrambling codes Eight sets

Figure 6.11: OVSF code tree.

Figure 6.12: Channelisation codes – – – code in use,


unavailable codes, ——— available

codes

Figure 6.13: Code allocation is branch sensitive (a) Codes C16 ; 8and C8 ; 4are on the same branch

and must not be used; (b) codes C16 ; 8and C8 ; 3are permissible

are formed (having 816=128 codes) to give a group of codes, and there are 64 groups.Each cell is allocated only one primary scrambling code The scrambling code hierarchy isshown in Figure 6.14

The block diagram for the down-link spreading and modulation is shown in Figure 6.15.The two outputs from the serial-to-parallel converter are multiplied by the channelisation

code, C ch , followed by multiplication by the complex scrambling code C scramb The output

of each multiplier is low pass filtered (LPF) by filters having a raised cosine impulse sponse with a frequency roll-off of 0.22 They are then multiplied by the quadrature carrierscos(ωc t)and sin(ωc t)and the quadrature signals added to give the transmitted carrier sig-nal, whereωc is the angular carrier frequency The arrangement is quadrature phase shiftkeying (QPSK) modulation

that provides a common demodulation reference over all or part of a cell; the primary mon control physical channel (P-CCPCH) that carries general network information; thesecondary common control physical channel (S-CCPCH) for paging and packet data; the

com-Figure 6.14: Scrambling code hierarchy.

Figure 6.15: Down-link spreading and modulation.

synchronisation channel (SCH) that a UE uses in its initial cell search; and the acquisitionindication channel (AICH) that controls the use of common up-link channels.

It has a predefined bit sequence, which for a single transmit antenna is an all logical 1sequence There are two types of CPICH: primary and secondary The P-CPICH provides

a coherent reference to obtain the SCH, P-CCPCH, AICH and PICH at the UEs, as thesechannels do not carry their own pilot information The channelisation code used by the

P-CPICH is C ch; 256 ; 0, an all logical 1 code, while its scrambling code is the cell’s primaryscrambling code In the case of a single transmit antenna, the CPICH is the unmodulatedprimary scrambling code There is only one P-CPICH in each cell, and is broadcast overthe entire cell

A UE estimates the channel impulse response from the received pilot signal, and armedwith this response the data may be recovered The procedure for achieving this is described

in Section 4.2.2.6 Thus the pilot and data must be transmitted over the same radio channel(which includes the transmitter and receiver antennae) Consequently, since the CPICH istransmitted over the entire cell or sector, it cannot be used to recover data from a narrowbeam of a smart antenna because the radio channels for the data and CPICH may be verydifferent A smart antenna with its narrow beams will create radio channels with few or nosignificant multipath components, unlike a wide angle beam

The secondary common pilot channel (S-CPICH) provides a common coherent referencewithin part of a cell or sector The antennae have narrow beams, e.g from a smart antenna,and may be used to target individual UEs or groups of UEs in close proximity to one another.The node B (a BS) may use any channelisation code having a length of 256 chips The S-CPICH may be used as reference for the S-CCPCH (which transmits paging messages) andthe down-link dedicated channels

the next down-link common physical channels on our list, namely the common controlphysical channels (CCPCH) The P-CCPCH is a 30 kb/s down-link channel that carries thebroadcast control channel (BCH) It is transmitted continuously over the entire cell Thespreading factor of the P-CCPCH is 256, and it is always scrambled by the cell’s primaryscrambling code The P-CCPCH does not have any TPC, TFCI or pilot bits It occupies90% of a slot as shown in Figure 6.16 (The first 256 chips, marked no transmission of theP-CCPCH, carry the synchronisation channel, a channel that occupies only 10% of a slot.)Observe that the P-CCPCH is transmitted in every slot, although there will generally bedifferent BCH data in each slot

The S-CCPCH transmits paging messages as and when required Data rates range from

Figure 6.16: Slot and frame structure of the primary common control physical channel.

30 kb/s to 1920 kb/s, corresponding to spreading factors of 256 to 4, respectively The datarate can be varied as required using the TFCI The S-CCPCH is scrambled by either theprimary or secondary scrambling code, and can be transmitted with or without its own pilotbits Packet data may be sent to particular UEs via the S-CCPCH The slot and frame struc-tures for the S-CCPCH are displayed in Figure 6.17 The S-CCPCH occupies a completeslot, and is transmitted in every slot However, the S-CCPCH is only transmitted when there

is information to send, unlike the continuous transmission of the P-CCPCH

paging channel (PCH) on an S-CCPCH The PICH carries page indicators (PIs), where a PIindicates to a subset of UEs within a cell whether they should examine the next S-CCPCHframe for paging messages Each UE is associated with a particular PI The PICH uses aspreading factor of 256, and its 10 ms frame has 15 slots (i.e the basic structure) Since eachslot contains 20 data bits, there are 300 bits per frame The PICH uses only 288 bits, leaving

12 bits unused In one frame the PICH may carry 18, 36, 72 or 144 PIs, which means thateach PI consists of 16, 8, 4 or 2 bits, respectively The PIs occupy a fixed position within

a frame, depending on the number of PIs per frame Figure 6.18 is an example when thePICH carries 18 PIs per frame, with 16-bit PIs All 16 bits will be set to a logical 1 if the PI

is set, i.e a paging message is being sent on the PCH, otherwise all the 16-bit PIs containlogical 0s

Ndata bits

Slot 0

Frame 0

Slot 1

Frame 1

Slot i

Frame i

Slot 14

Frame 71

Frame 0

Slot 1

Frame 1

Slot i

Frame i

Slot 14

Frame 71

Figure 6.18: Page indication channel frame showing an example of 18 page indicators.

The synchronisation channel A cardinal function in any communication system is chronisation In UTRA the synchronisation channel (SCH) is used by UEs in the initialcell search process In Figure 6.16 the primary CCPCH occupies 90% of each slot TheSCH is transmitted in the first 10% of these slots, during 66:67 µs by a 256-chip word.

syn-There are two sub-channels of the SCH, the primary and secondary synchronisation nels Figure 6.19 shows the primary and secondary SCH over a 15-slot frame The primary

chan-synchronisation code (PSC) (shown as C pin the figure) is the same in every slot, and is thesame in every node B, although the node Bs are not synchronised The PSC is a generalisedhierarchical Golay sequence, chosen to reduce the complexity of the cell search procedure

It has good autocorrelation properties as can be seen with reference to Figure 6.20

The PSC is generated by modulating a 16-chip code running at 3.84 Mchips/s by another16-chip code generated at 240 kchips/s The result is a 256-chip sequence at 3.84 Mchips/swhose autocorrelation function can be rapidly found, as we will describe in the cell searchprocedure A UE receiver utilises the PSC initially to detect the presence of a nearby BSand then to identify the start of each timeslot

The secondary SCH (shown as C i; 1

s , j=0;1; : ;14 in Figure 6.19) is transmitted taneously with the primary SCH during the first 256 chips of every slot The BS transmitterarrangement is shown in Figure 6.21 The basic function of the secondary SCH is to enablethe identity of the scrambling code group used by the BS to be determined by a UE Weobserve in Figure 6.19 that there are 15 different 256-chip secondary synchronisation codes(SSCs), i.e the SSC changes from slot to slot such that the SSCs in a frame constitute a

simul-predefined sequence that is associated with the scrambling code group used by the cell; see

Figure 6.14 Figure 6.22 shows an example of the SSC sequence for scrambling code group1

There are 64 scrambling code groups, and with the aid of the SSC a UE knows the actualgroup number The group identified has eight sets, and each set has a unique primaryscrambling code This code is also the pilot code The UE receiver cross-correlates thepilot code with all the eight primary scrambling codes of the eight sets in the group Bythis means the receiver determines the correct primary scrambling code Since this codealso scrambled the BCH data, this data can now be recovered The main goal of the SSC

is thereby achieved The BCH data, include, if appropriate, the secondary scrambling codebeing used The secondary scrambling codes are used when spot beams are employed

The PSC, Cp, and the SSC sequence C i;j

s ;i=1; : ;512 and j=0;1; : ;14, are added to

the other down-link channels The Cp and C i;j

s are not subjected to multiplication by eitherchannelisation or scrambling codes

acquisition indicators (AIs) An AI is part of the random access procedure which indicates

Cp Cp Cp

· · ·

256 chips

Figure 6.20: Primary synchronisation code autocorrelation function.

Figure 6.21: BS transmitter block diagram.

16 7 15 7 2 16 10 8 15 10 9 8 2 1 1

14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

16 7 15 7 2 16 10 8 15 10 9 8 2 1 1

14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

Slot Number

Csi,0 Csi,1 · · · Csi,14

256 chips

10 ms frame

Csi,2

Scrambling Code Number

Figure 6.22: The SSC sequence for scrambling code group 1.

to a UE that it should proceed with the transmission of its access request The AI can takethree different values: a 0, which causes the UE to transmit with more power until it gets

a 1 or +1 response; a 1, which instructs the UE to discontinue with the access attempt;and a +1 which enables the UE to go to the next stage in the access procedure Observe thatthe three states of 0, 1, and +1 result in no code being sent, the code sent but inverted, andthe code sent, respectively

The AICH has a 15-slot frame, whose duration is 20 ms rather than the usual 10 ms.Accordingly each slot has a duration of 1.33 ms, equivalent to 5120 chips The first 4096chips correspond to 16 AI symbols as the spreading factor is 256 The last 1024 chips arefour zero AI symbols Figure 6.23 shows the slot and framing arrangement

The AI part of the slot has 16 orthogonal code words which have been multiplied by the

AI Each code word corresponds to one of 16 different Walsh codes, each of 256 chips

These codes are called preamble signature codes (We will discuss these codes later when

we come to the access procedures.) Figure 6.24 shows the 16 preamble signature codes

W1;W2; : ;W16 Each code will be associated with a particular UE, and this code is tiplied by the appropriate AI, e.g W2 AI2, resulting in either no output, or all the chipsinverted in W2, or just W2, depending on whether AI2 is 0, 1 or +1, respectively Up

mul-to 16 WiAIi;i=1;2; : ;16, are summed and loaded into the AICH slot The switch ting in Figure 6.24 for the last four AI symbols introduces logical 0s, which means that nocodes are transmitted Consequently, the AICH slot has data corresponding to 16 symbolsarranged over the first 80% of the slot

down-link transport channels that define the manner in which information is mapped ontothe physical channels There are two categories of transport channels: the dedicated channel(DCH), and the common transport channels

There is only one DCH, and it has the ability rapidly to change the transmitted data rate,

if necessary, every 10 ms It also supports fast power control, and provides addressing ofUEs The DCH is always mapped onto the DPDCH, and may be transmitted over the entirecell, or to zones illuminated by narrow beam antennae

The situation is more complex for common transport channels The broadcast channel(BCH) is a point-to-multipoint channel that is used to broadcast network and cell-specificinformation to all UEs within a cell It is mapped onto the P-CCPCH The forward accesschannel (FACH) carries small amounts of information such as short packets of user data.The FACH does not need to be transmitted over the entire cell, e.g if smart antennae areused, and it is mapped onto the S-CCPCH The paging channel (PCH) is transmitted overthe entire cell and conveys pages to the UEs It uses the S-CCPCH